How to Automate Your Workday with AI (Step-by-Step Guide)

If you want to automate your workday with AI, the goal is usually the same: stop doing the same tasks over and over. Drafting the same types of emails. Summarizing documents they barely have time to read. Moving information from one app to another. These aren’t big tasks — but according to Asana’s Anatomy of Work Index, knowledge workers spend about 60% of their day on exactly this kind of repetitive busywork rather than the skilled work they were actually hired to do.

This guide covers how to actually automate your workday with AI — not the theoretical version, but the practical one. The kind that works for a regular person who uses email, has meetings, writes things, and doesn’t want to learn to code. I’ll walk through the tools that fit this use case, the specific automations worth building first, and the ones that sound good but rarely pay off in practice.

New to AI tools? Start with Best AI Tools for Beginners: Start Free, Upgrade When It’s Worth It before building automations — it’ll save you time choosing the right starting point.

Step 1: Figure Out What’s Actually Worth Automating

Before picking tools, spend five minutes answering one question: what do I do manually at least three times a week that I hate doing? That’s your automation list. Not everything is worth the setup time, and the biggest mistake people make is automating something they only do occasionally. The opportunity is real: McKinsey’s 2025 research found that a significant share of current work hours could potentially be automated with today’s technologies — but most people are still doing them manually, one task at a time.

Tasks that are almost always worth automating:

• Writing routine emails (responses, follow-ups, status updates)
• Summarizing meeting notes or long documents
• Moving data between apps manually (copy-pasting from one tool to another)
• Writing first drafts of reports, updates, or recurring content
• Sending the same type of message to multiple people

Tasks that are usually not worth automating right now:

• Anything that requires real judgment or stakeholder nuance
• Tasks you do less than once a week
• Complex multi-step workflows where one error cascades
• Tasks where the output genuinely needs to sound like you — not a template

Step 2: Automate Email Drafting

Email is the highest-ROI place to start for most people. Most of us write the same types of emails on repeat: follow-ups, status updates, requests, thank-yous, polite declines. These are predictable in structure, so AI handles them well — and the time savings compound fast if you’re sending 20+ emails a day.

Option A — ChatGPT (free, manual)

The simplest approach: keep a browser tab open with ChatGPT and use a standing prompt template. When you need to write an email, paste your context and get a draft in seconds.

Prompt template to save and reuse:

I used this to draft a follow-up email after a client call that had gone sideways — gave ChatGPT the bullet points of what was agreed, the tone I wanted (professional but direct), and had a send-ready draft in under two minutes. Editing took another 30 seconds.

Option B — Gmail’s built-in AI (free, integrated)

If you use Gmail, “Help me write” is already in your compose window — it’s available on personal Gmail accounts and Google Workspace alike. Type a brief description of what the email needs to say and Gmail drafts it inline — no tab switching. For straightforward emails, this is the lowest-friction option available. The quality is solid for simple tasks, though it doesn’t match ChatGPT’s output for nuanced or longer emails.

Option C — Jasper AI (paid, brand-consistent)

If you’re producing high-volume marketing or business emails that need to sound consistently on-brand, Jasper’s email templates and Brand Voice feature handle this better than ChatGPT’s blank canvas. Worth considering once ChatGPT has proven the use case for you. See the full Jasper AI review →

Step 3: Automate Meeting Notes and Follow-Ups

Meeting notes are one of the most universally dreaded tasks in office work — everyone has to do them, almost nobody does them well under time pressure, and they pile up fast. AI has made this one of the easiest workday tasks to automate, and the tools available now are genuinely good.

Built-in AI in your meeting tool

Most major meeting platforms now include AI transcription and summarization. Check what’s already available in the tools your team uses:

• Google Meet — Gemini AI summaries on Workspace Business and higher plans. Check current plan availability →
• Microsoft Teams — Copilot transcription and summaries on Microsoft 365 Copilot plans. See Microsoft 365 Copilot details →
• Zoom — AI Companion summaries included on paid plans. Confirm your plan includes it at zoom.us/pricing →

If your meeting tool already has this, enable it first before adding a third-party app — it’s already paid for and requires zero additional setup.

Standalone AI meeting assistants

If your meeting tool doesn’t include AI, or you want more control over how notes are structured, dedicated meeting assistants fill this gap well. Tools like Otter.ai, Fireflies.ai, and Fathom join your calls as a bot, transcribe in real time, and send a structured summary with action items when the call ends. Most have a free tier that covers a meaningful number of meetings per month — enough to test whether the workflow actually fits how you run calls. I’ve been using Fathom for calls where I need clean action items without manual cleanup — the summary it sends after the call is usually 80–90% usable as-is, which is what sold me on it.

Manual fallback: ChatGPT + your rough notes

If you already take rough notes during calls, paste them into ChatGPT with this prompt: “Clean these up into a structured meeting summary with key decisions and action items.” It takes messy bullet points and produces a shareable summary in seconds. Not fully automated, but the editing time drops to almost nothing.

Step 4: Automate Document Summarization

Long reports, lengthy email threads, policy documents, research papers — the amount of reading most knowledge workers have to do is genuinely unsustainable. AI summarization doesn’t replace reading when you need to, but it’s very good at helping you triage: figuring out which documents actually need your full attention and which can be handled with a three-paragraph summary.

ChatGPT for one-off document summaries

Paste the text of a document into ChatGPT (or upload a PDF in ChatGPT Plus) and ask: “Summarize this in 3 bullet points: the main argument, the key data point, and what action is being requested.” For most professional documents, this gives you exactly what you need in under 30 seconds.

Notion AI for your ongoing document library

If your work documents live in Notion, the AI layer is more useful than a standalone summarizer because it works within your existing workspace. You can highlight a page or block and ask for a summary, action items, or a rewrite — without leaving the app. Worth noting: Notion AI summarizes the specific pages or content you point it to; it doesn’t automatically search across your entire workspace unless you explicitly ask it to within a given page. As of 2025, full Notion AI access requires the Business plan ($20/user/month) — the separate add-on has been discontinued for new users. Pricing and plan structure are subject to change — confirm current options at notion.com/pricing before purchasing.

Google NotebookLM for research-heavy documents

If you regularly work with long PDFs, research reports, or dense reference material, Google’s NotebookLM is worth knowing about. You upload your documents and it lets you ask questions directly against them — grounded in the text, with cited sources. It’s free to use and handles multi-document projects well. Useful for anyone who has to synthesize information across several files at once rather than summarize a single document.

Claude.ai for longer or more sensitive documents

Claude (by Anthropic) handles very long documents well — its context window (the amount of text it can process at once) is one of the largest available, which matters when a report runs to 50+ pages and you don’t want to split it into chunks. For users on business or team plans, data privacy protections are also stronger than the consumer ChatGPT default, which makes it a reasonable choice for documents you’d rather not paste into a shared platform.

Which tool to choose: If you need to summarize a one-off document quickly, ChatGPT is the fastest path. If your documents live in Notion already, use Notion AI — no copy-pasting needed. If you’re working across multiple PDFs or research files, NotebookLM handles multi-source queries better than any of the others. And if a document is long (50+ pages) or contains sensitive information, Claude is the most practical option.

Once you’ve got document triage working, the next level is cutting out the manual work that happens between your apps entirely.

Step 5: Automate Repetitive Workflows Between Apps

Think about how many times a day you copy something from one app and paste it into another. A form submission goes into a spreadsheet. A new client goes into your CRM. A completed task needs a Slack message. Each one takes 30 seconds — and across a full day, those 30 seconds add up to a surprising amount of lost time. This is exactly the gap Zapier was built to close.

Zapier connects thousands of apps and lets you build automations — called Zaps — through a visual interface with no code required. A Zap has two parts: a trigger (something that happens in one app) and an action (something that happens automatically in another app as a result).

Five Zaps worth building first

Start with Zap #1 or #2 — they’re the fastest to set up and immediately eliminate a daily friction point. The first Zap I built was a version of #2: new form submissions from Typeform logging straight into a Google Sheet and firing a Slack notification. Setup took about 25 minutes including testing. By the end of the first week it had run 40+ times without me touching it once. Here’s what to expect from Zapier’s plans as your needs grow:

Prices and plan details are subject to change at Zapier’s discretion. — verify current pricing directly at zapier.com/pricing →

Step 6: Automate Content and Writing Tasks

If part of your workday involves producing written content — blog posts, social media updates, internal reports, product descriptions — AI handles first drafts better than almost anything else in this guide. The key is treating AI output as a starting point, not a finished product. The blank page problem disappears; the editing and judgment stays with you.

I use this most for recurring deliverables — the weekly status update I send to stakeholders, and repurposing longer content into shorter formats for different channels. What used to take 30–40 minutes now takes about five, mostly spent on light editing. The prompts below are the ones I’ve refined over time and actually reuse.

Prompts that actually work for recurring content

Weekly status update:

Social media post from a blog post or article:

Internal project update or brief report:

Save your most-used prompts in a Notion doc or a pinned ChatGPT conversation — they’re reusable and get better the more you refine them over time. This turns a 30-minute writing task into a 5-minute editing task.

Those six steps cover the automations most worth your time. Before you go further, it’s worth knowing which ones consistently disappoint — so you don’t waste a Saturday setting something up that won’t pay off.

What’s Not Worth Automating (Yet)

Not every automation idea is worth the setup cost. These four come up most often — and consistently underdeliver. I’ve tried most of them myself and ended up reverting or abandoning them within a few weeks.

What Do These Tools Actually Cost?

Most tools in this guide have a usable free tier. Here’s a quick reference — always verify current pricing at each tool’s official site before committing to a paid plan. Best value for most people starting out: ChatGPT free + Fathom free covers email drafting and meeting notes at zero cost.

Prices are estimates and change frequently — always confirm at the official site before purchasing.

How to Get Started This Week

The worst thing you can do is try to automate everything at once. Here’s a realistic three-step plan for this week:

By the end of the week, you’ll have three automations running — and a clear sense of which ones actually changed how your day feels. Start there before building anything more complex.

Frequently Asked Questions

Automation doesn’t have to be a big project. The people who get the most out of it start with one small thing, prove it works, and build from there. One prompt. One meeting tool enabled. One Zap. That’s enough to change how your week feels — and it’s enough to know whether more is worth it.

What’s Next

Once you’ve got the basics running, these are the most useful next reads for building out your AI productivity stack:

✍️ We test and use AI tools in our own workflows — no jargon, just honest guidance based on real experience. About DailyTechEdge →